Shopping Cart Commodity Recommendation via Image Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current supermarket shopping models are inefficient as users need to manually search for commodities, leading to high time costs due to commodities being arranged only by categories.
Innovation Solution
A commodity recommendation method utilizing a server and shopping cart system that collects images of commodities, performs recognition and recommendation using neural networks and correlation matrices to suggest relevant items based on user preferences and shopping history.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If commodities are arranged only by categories in supermarkets, then the arrangement is simple and easy to manage, but users need to manually search for commodities leading to high time costs
Solution Approach 1:
The shopping cart system automatically captures images of commodities, performs recognition and recommendation without user intervention. The system serves itself by autonomously identifying commodities in the cart and generating recommendations, eliminating the need for users to manually search for products.
Solution Approach 2:
The patent replaces the mechanical manual searching process with an automated image recognition system using convolutional neural networks. The system captures images via camera, processes them through neural network models to identify commodities, and generates recommendations automatically, substituting human manual search behavior with automated computational processing.
2Measurement precision
If multiple images of the same commodity from different viewing angles are collected, then recognition accuracy is improved, but the amount of data to be processed increases
Solution Approach 1:
The patent segments the image processing task by using multiple convolutional neural network models, each specialized in recognizing commodities from specific viewing angles. This segmentation allows the system to process multi-angle images more efficiently by distributing the recognition task across specialized models rather than using a single general-purpose model.
Solution Approach 2:
The system performs preliminary classification of images by viewing angle before recognition. By pre-categorizing images according to their viewing angles and routing them to appropriate specialized neural network models, the system prepares and organizes data in advance, reducing the complexity of the actual recognition process.
Data Source
AI summary
A commodity recommendation method includes: receiving at least one image of a first commodity sent by a shopping cart, and the first commodity being a commodity placed in the shopping cart; inputting the at least one image into a commodity recognition model for commodity recognition to obtain commodity recognition information; inputting the commodity recognition information into a commodity recommendation model for purchase demand prediction to obtain at least one target recommended commodity; and sending information of the at least one target recommended commodity to the shopping cart.